Legion Theorycrafting [7.3, All specs]

[Effinhunter]

Of course. Discussion is encouraged. So is posting your own sims, results, and theorycrafting. I'm still traveling for work, but if there are specific sims needed, then I can help out.

[bouchbagfett]

I used AskMrRobot sims since it is easy to do from work but my question is how good is the Iskar mode for estimating AoE pull dps for Bestial Fury/Blink Strikes and Stampede/Killer cobra and Thunder Strikes/Furious Strikes relics.

I guess the question to me is how much single target vs AoE we should go for and what will get more use, Stampede or Killer Cobra. Stampede is more powerful per use, especially AoE but limited by the CD. Killer Cobra has great single target but when you are doing AoE packs Kill Comand usage goes way down.

Bestial Fury/Killer Cobra/Thunder Strike have good synergy. Blink Strikes/Furious Strikes also have good synergy. To me it is not so clear what will be most optimal for damage and success.

My MrRobot sims show a 8.7% gain in single target DPS by going Bestial Fury/Killer Cobra/Thunder Strike over Blink Strikes/Stampede/Furious Strikes but the later does 12% more AoE damage than the former.

[GT4]

I haven't gotten to play MM yet on my hunter alt (got 2nd major for mythic+ first) so I haven't gotten to refine the rotation yet either but it seems like the default simc rotation isn't quite optimal, at least for the default profiles. E.g. adjusting the line
actions+=/sidewinders,if=debuff.hunters_mark.down&(buff.marking_targets.remains>6|buff.trueshot.reac t|charges=2)
to
actions+=/sidewinders,if=debuff.hunters_mark.down&(buff.marking_targets.up|buff.trueshot.react|charg es=2)
nets a gain of a few hundred DPS for each default profile. I wouldn't be surprised if a few other small tweaks would yield similar returns.
In addition to that, the default APL doesn't seem to use aimed shot without vulnerability at all. Have you tested whether low haste setups might be better off dumping without vulnerability up in certain situations?

[Effinhunter]

I used AskMrRobot sims since it is easy to do from work but my question is how good is the Iskar mode for estimating AoE pull dps for Bestial Fury/Blink Strikes and Stampede/Killer cobra and Thunder Strikes/Furious Strikes relics.

I guess the question to me is how much single target vs AoE we should go for and what will get more use, Stampede or Killer Cobra. Stampede is more powerful per use, especially AoE but limited by the CD. Killer Cobra has great single target but when you are doing AoE packs Kill Comand usage goes way down.

Bestial Fury/Killer Cobra/Thunder Strike have good synergy. Blink Strikes/Furious Strikes also have good synergy. To me it is not so clear what will be most optimal for damage and success.

My MrRobot sims show a 8.7% gain in single target DPS by going Bestial Fury/Killer Cobra/Thunder Strike over Blink Strikes/Stampede/Furious Strikes but the later does 12% more AoE damage than the former.

If they're modeling Iskar, then you have add packs that are appearing at 70%, 45% and 20% health. This is atypical of most add fights, and is going to give you very mixed results if you try to apply it to general AoE. One of the problems with using a default APL in this kind of fight is that the APL will use all cds at the top and will not save CDs for the add phases. For Mythic Iskar, towards the end of the expansion, there were several different strats including 1) burn down the phantasmal before he can chains, 2) burn the boss to get two add phases back-to-back, and 3) combinations of cds used per add wave, depending on dps comp and such. Testing for these strats requires custom APLs, if you want to do this right. It's certainly possible, but I doubt that was being done in your sims. You really have to decide where/when dps needs to be applied to do these custom APLs. Certain strats/bosses/etc. require add phases to be handled as quickly as possible. Some require more boss focus.

- - - Updated - - -

I haven't gotten to play MM yet on my hunter alt (got 2nd major for mythic+ first) so I haven't gotten to refine the rotation yet either but it seems like the default simc rotation isn't quite optimal, at least for the default profiles. E.g. adjusting the line
actions+=/sidewinders,if=debuff.hunters_mark.down&(buff.marking_targets.remains>6|buff.trueshot.reac t|charges=2)
to
actions+=/sidewinders,if=debuff.hunters_mark.down&(buff.marking_targets.up|buff.trueshot.react|charg es=2)
nets a gain of a few hundred DPS for each default profile. I wouldn't be surprised if a few other small tweaks would yield similar returns.
In addition to that, the default APL doesn't seem to use aimed shot without vulnerability at all. Have you tested whether low haste setups might be better off dumping without vulnerability up in certain situations?

I don't doubt this at all. I need to reinvestigate the APLs of all three specs soonish. Unfortunately, I have a student PhD defense next week and I have to catch up on everything I've missed during my gaming PTO and travel. I'll try to make time this weekend.

Were you curious about the stock build specifically, or are you trying to use AiS without vulnerability with a different talent configuration?

[Zoopercat]

One of the problems with using a default APL in this kind of fight is that the APL will use all cds at the top and will not save CDs for the add phases.

Actually, we DO check for add phases in our rotations, in AMR.

We have a few functions that you can see in our default rotation that allow for AoE saving that works across all fights. Here's the default rotation on the open-source, public wiki: http://www.askmrrobot.com/wow/theory...p&version=live (See the screenshot below for the areas that check for saving AoE).

Iskaar is a pretty good AoE test. I'd argue it's much better than a script that has 100% AoE all of the time, which never happens.



The AMR simulator is really powerful with built-in functions that make rotations really smooth. And while we only have Iskaar and Zakuun modeled right now, in a week or two, we'll have the EN bosses so you can do theory against real, current fights

[bouchbagfett]

And while we only have Iskaar and Zakuun modeled right now, in a week or two, we'll have the EN bosses so you can do theory against real, current fights

What about simulated 4ish mob stead trash pulls to Mythic+ boss?

BTW really enjoying AMR sim tool.

[Effinhunter]

Actually, we DO check for add phases in our rotations, in AMR.

We have a few functions that you can see in our default rotation that allow for AoE saving that works across all fights. Here's the default rotation on the open-source, public wiki: http://www.askmrrobot.com/wow/theory...p&version=live (See the screenshot below for the areas that check for saving AoE).
...
The AMR simulator is really powerful with built-in functions that make rotations really smooth. And while we only have Iskaar and Zakuun modeled right now, in a week or two, we'll have the EN bosses so you can do theory against real, current fights

I'm not arguing that you can't make APLs that cater to the fight, only that by default, most APLs do not perform the fight using the tactics you might actually use. For instance, I can't tell from your screencap when Trueshot is being used, but saving it for the Iskar add phase will greatly benefit the strats that call for burning down all the adds quickly, especially the Phantasm who casts the chains. If it's used with Bloodlust at the top, it's best for certain other strats--like the one that burns in a single ring to unleash Maalus on the add or the boss, depending on the strat, to push him into a chained 2nd add phase or whatever. In this APL case, you're not saving lust for the add phase, right? You're using it at the top, and most APLs would use Trueshot with Bloodlust at the top, meaning your add dps will be less, but when we had the legendary ring, especially, using everything at the top with trinket procs was obviously a really great idea for overall dps.

Generally, when trying to make an accurate boss fight script in Simcraft, I would also tweak the APL to save cds for add phases, for instance, if that's what the fight called for, or not, if boss dps was actually what should have been prioritized. It usually takes a lot of work to get a boss script, especially one with adds, to really come close to the real thing.

But Iskar reflecting trash dps? I assumed this guy was asking about AoE dps sims for Mythic+ dungeon trash pulls. If he's wanting to use it for boss fights, then he'd probably need to tweak the default APL to better reflect whether he wants to maximize the AoE component. I think the AMR team is doing great work, but creating a perfect default APL for each boss fight is not only really hard to do, but it is usually going to depend on a guild's strategy for defeating a particular boss. And most guilds change boss strategies over time as they get better gear (i.e., farm dps opportunities are usually very different from progression).

- - - Updated - - -

What about simulated 4ish mob stead trash pulls to Mythic+ boss?

BTW really enjoying AMR sim tool.

Case in point, Zooper!

At bouchbagfett, I use staggered add sims for this. All trash packs in all dungeons have high health mobs and low health mobs. Simming this correctly requires that some of the mobs die off faster than others, but it's going to depend on group comp as to how accurately you can model this. For instance, if you are running with 2 havoc DHs, they have really great initial add burst that often kills the smaller mobs so quickly that a 6 target sim quickly turns into a 2 target sim (I run with 2 really good DHs frequently). You can see examples of staggered adds in the first post under the Mythic+ discussion. I'm sure AMR has a similar feature.

- - - Updated - - -

P.S. If there are some good hunters with 13 traits of BM on the US-side that would be willing to run some dungeons with me, I'd like to see how close 13 trait BM is actually coming to MM. I have run with 12+ 845-849 ilvl 13 trait BM hunters in pugs and various other dungeons against 16-20 trait MM, and I have not seen BM beating MM like some are claiming it does. Our hunters have pretty much given up on 13 trait BM beating MM at this point (we've seriously tried very hard to make this work, even when pulling 10+ mobs, but the guaranteed Marked Shot proc on pull means that MM has some ridiculous burst as well and doesn't have to wait for Stampede). If you think you are pushing what is possible in BM, please send me a message if you have some time to test things out. I would appreciate it. Effinhunter#1904.

[eathdemon]

I assume yes, but want to double check, is +115 agi and + 500 haste worth the loss of 200 mastery

[Finbezz]

I assume yes, but want to double check, is +115 agi and + 500 haste worth the loss of 200 mastery

uh, yea, not even close.

[Zoopercat]

@Effinhunter ah yes, mythic+ we don't have a script for yet.

If you click on that link I gave you can see the whole rotation. I'm sure people will customize them to each boss once we have all of those coded in. What I meant to really point out was our 'save for aoe' functions that make it pretty easy to implement

I still have to fill you in on some of the hardcore features when we both have some time.

[Jayeazy]

Not to turn this into an amr thread but since we're talking about it, how can I change apl shit like this.



I see no way to edit it even clicking view/edit.

[Zoopercat]

@TheJazz to edit it, you need to log in, because we need an account to save your changes to. Once you login, you'll see a 'copy' button which then lets you edit your own version.

[Jayeazy]

Oh well that's easy, thanks Zooper.

I was bored and worked up a new mm sw spec apl from scratch because though the apl as it was gave pretty good results, I thought it could do a few things better. I'm still tweaking it and have yet to fit any other talents than what were in the default profile, but the results are decent so far.

ST is seeing a small increase of like 1k so far:

300 sec +/- 20% at 100k

Since it's so much different I can't exactly say how the behavior changes, but I can explain some of the things I saw happening in the default profile I disagreed with. In the default apl, under Trueshot, it seemed to worry more about dumping focus with Aimed Shot in the beginning so it wasn't focus capped for the first Sidewinders. That delays that first sidewinders by enough to lose a 4th Marked Shot under Trueshot + Hero. Changing the priority to SW>MS>Ai>etc while under Trueshot gave a few hundred dps consistently. It also backs up what Azor's guide advises to open with. I also simplified the conditions of using Wwindburst. I saw many people saying to use Windburst as a way to artificially extend a SW/MS cycle and would only use it under the condition that Vuln is down or falling down as it's cast. I saw virtually no difference in sims when changing it to simply being used on cooldown as long as Marked Shot can be cast after it (time/focus check) if Hunters Mark is up. In practice, it would also be much easier for players to handle like that. That said, it can definitely still be a dps gain on top of a dps loss if you need to use it to remedy a bad situation due to something like excessive movement, you just need to be smart about your situation I guess. Also, Windburst was completely ignored in the default APL aoe, so that's back in.

3 target saw more noticeable changes:


and 5 target:


both at 300 sec +/- 20% at 10k

One thing I had noticed was that the priority target damage from the default profile wasn't exactly portraying just how good mm's ability to cleave with hardly any sacrifice on ST is. To explain some of the labels:

"ms prio for mt" - means that Marked Shot is used as soon as possible on multitarget

"mt aware" - is basically a workaround for a Vuln timing problem. Since on ST it's possible to extend Vulnerable with Windburst while Hunters Mark is up on your main target, the sim would only be concerned with getting the Marked Shot off while that main target's Vuln was still up, so it was possible that Vuln fell off secondary targets before Marked Shot went out. Any sim with this label will target a secondary target when checking for a good time to Marked Shot so that all secondary targets are hit by it while Vuln is up on all of them.

So obviously the new ST profile was the highest prio damage because it made no sacrifices whatsoever to maximize aoe, yet was still very good and beats the default aoe up until a few conditions kick in for it in the 5 target sim.

Being "multi target aware" gives good gains obviously since it tries to make sure all Marked Shots are buffed on all targets.

Prioritizing Marked Shot on any number of targets gave the best gains but at the biggest loss to main target damage. However, comparing the default and new apl, more aoe damage is gained and less main target damage is lost with the new apl.

Combination of a ms prio with being mt aware resulted in no change obviously since if marked shot is being cast so early, there isn't a chance for the sim to lose track of it based on the main target.

Overall, I think the new apl with a ms prio for multi target is the most efficient way to go since the main target loss is minor but sees the most aoe gains. That said, simply being multitarget aware with the new apl loses no main target damage and sees a boost in aoe damage. In practice, that just has to be decided based on the situation.

http://pastebin.com/KhEX9mwC

Anyway, that's the profile I'm using and like I said, I'm still tweaking things and finding oddities (over 3 second Barrage cast times...) but it's definitely much different and relies on more conditionals than the default apl.

[GT4]

I don't doubt this at all. I need to reinvestigate the APLs of all three specs soonish. Unfortunately, I have a student PhD defense next week and I have to catch up on everything I've missed during my gaming PTO and travel. I'll try to make time this weekend.

Were you curious about the stock build specifically, or are you trying to use AiS without vulnerability with a different talent configuration?

I'm just messing around as of now. Just wanted to mention this here as it might affect stat weightings (e.g. haste being overvalued because the rotation is more ideal for high haste gear where vulnerability has higher uptime).

I've also had a suspicion regarding SV yesterday and I've confirmed it today: Fury of the Eagle damage is completely off in simc.

In simc, importing my character and replacing the APL to do nothing but use auto attacks and fury of the eagle with bloodlust disabled and looking at the log, this is what I noticed:
1. In simc, it does 1 initial hit and then 9 ticks for 10 hits in total. In-game, it only does 1 initial hit and then 8 ticks for 9 hits in total.
2. In simc, the initial hit does 20081 damage and the ticks do 39080 damage each. In-game, the initial hit and each tick do 18999 damage each.
3. In simc, auto attacks (pet and hunter) seem to have no damage range unlike in-game (not necessarily important in itself but becomes important for the next point).
4. In simc, the initial hit does exactly the same damage as an auto attack and the ticks roughly double that damage. Assuming this is no coincidence and it's actually calculated based on weapon damage (as is also indicated by the report which says weapon multiplier = 1.00), this contradicts to the in-game tooltip which only scales off of AP (and is also wrong by the way, see below).

Now looking at in-game data again and comparing different AP levels, the actual damage is 85% of my AP per tick (or 765% of my AP total) at each AP level which means damage is indeed a multiplier of AP and not a multiplier of weapon damage as apparently simc currently uses.

Going by my pet's auto attacks (which have no damage range in-game either), damage in simc and in-game exactly matches so there don't seem to be any greater issues. Thus I'll just assume the damage multiplier from armor in simc is correct at ~0.68 which means FotE should be doing 0.85 / 0.68 ~= 1.25 times your AP per hit for a total of 1.25 * 9 = 11.25 times your AP (1125%). The in-game tooltip currently shows 625% of AP which is equal to 5 * 125% so it seems like the in-game tooltip only takes into account 5 hits instead of the actual 9 hits.

tl;dr: In simc, Fury of the Eagle does 100% of your weapon damage initially and 9 ticks of ~200% of your weapon damage each for a total of 1900% of your weapon damage. In-game, however, Fury of the Eagle does 125% of your AP initially and 8 ticks of 125% of your AP each for a total of 1125% of your AP. Doing some quick napkin maths, this should almost halve FotE damage for a loss of ~9% total damage in simc, possibly altering stat weightings as well. It might also account for the difference between SV and MM/BM that's shown in simc but not in-game according to some people.

[Effinhunter]

@TheJazz and @GT4, I haven't had a chance to validate any of this, but I definitely appreciate the contributions. For the SV issues, this would explain the very weird jump we saw after the big expansion patch in Simcraft. If Fury of the Eagle is badly modeled, this will need to be fixed asap in the hunter module. Has anyone else verified this in game behavior?

[Azortharion]

[Deleted]

so does it do more or less dmg after the fix?

[GT4]

Damage certainly matches my in-game data now (e.g. for the previous test case, 63 ticks of 18999 damage for 7 uses). Going by the change on the commit, it seems like simc simply added weapon damage to the application and each tick (resulting in 10 times your weapon damage extra) because it parsed the spell data wrongly.

so does it do more or less dmg after the fix?

Way less. SV lost >20k DPS for the default profile and FotE went from >19% of total damage to 8% of total damage.

Edit: Did a quick check when it comes to stat values for SV. Highest DPS I was able to get by shifting stats around in the default profile was high versatility and a split between haste and crit for the remaining stats.

[Azortharion]

Something something Survival not really all that great for AoE or Mythic dungeons.

- - - Updated - - -

Btw good find, it definitely pays off to test ingame Survival vs SimC especially because its APL and module has not really been highly prioritized, much easier for bugs to linger. Especially with non-"standard" talents.

[Effinhunter]

All of us have been saying the SV damage looked too high. It even says it in bold letters in the first post, so this is hardly surprising. I'll be rerunning all sims for SV soon.